Humans walk upright, carry brains three times larger than those of our closest relatives, and possess slender, distinctly shaped skeletons. Yet pinning down exactly which DNA changes produced these features has remained one of biology’s most stubborn mysteries. Now, researchers at the Institute for the Advanced Study of Human Biology (ASHBi) at Kyoto University have delivered a major breakthrough: the first functional atlas of the regulatory changes that shaped the human skeleton, published in the journal Nature on 23 September 2026.

The puzzle begins with a striking numerical fact. Human and chimpanzee genomes are more than 98 percent identical, and the protein-coding genes themselves—the blueprints for our proteins—barely differ at all. This has long led scientists to conclude that the essence of what makes us human lies not in the genes themselves but in the vast stretches of non-coding DNA that act as genetic switches, determining when, where, and how strongly each gene is turned on. The technical challenge has been equally daunting: for any given sequence difference, it has been extraordinarily difficult to determine whether a human-specific change actually alters gene activity or is merely a harmless bystander.

To overcome this bottleneck, the research team focused on the human skeleton and combined two complementary experimental approaches on a genome-wide scale. First, they compared the genomes of more than 15,000 people with those of 139 great apes, including chimpanzees, bonobos, gorillas, and orangutans. This comparison identified roughly 5.7 million changes that arose only along the human lineage, of which about 561,000 fell within candidate promoter and enhancer regions—the regulatory sequences that control nearby genes.

The researchers then synthesized both the human and ancestral forms of each of these sequences and tested them in cultured cartilage cells using a technique called a massively parallel reporter assay, or MPRA. This method measures the activity of hundreds of thousands of regulatory sequences simultaneously, allowing the team to sift through the enormous candidate list with unprecedented speed. The screen narrowed the field to approximately 15,000 human-specific variants that measurably altered gene activity, transforming a catalogue of sequence differences into a functional map of evolutionary change.

The second approach addressed a different problem: how to test the effects of these regulatory changes in a shared cellular environment. The team fused human and gorilla induced pluripotent stem (iPS) cells to create interspecies hybrid cells, then directed them to develop into progenitor cells capable of forming bone and cartilage. Because the human and gorilla chromosomes occupied the very same cells, the researchers could compare their gene activity directly, eliminating the confounding effects that arise when the two species are studied in separate cultures. Applying the same strategy to previously generated human-chimpanzee hybrid cells, they identified genes that are regulated differently in humans compared with other great apes.

The strongest signal to emerge from this analysis pointed to pathways involving glycosaminoglycans, or GAGs. These long sugar molecules are major components of the extracellular matrix, the scaffold that surrounds and supports cells, and they help maintain the structure, elasticity, and water content of cartilage. In the human-gorilla hybrid cells, many GAG-related genes were consistently less active on the human chromosomes than on their gorilla counterparts—a pattern consistent with the effects of human-specific regulatory changes accumulated over millions of years of evolution.

Crucially, the same pattern appeared in real tissue. When the researchers measured GAG content in elbow and knee samples from humans and other great apes, levels varied little among the other apes but were only about one-quarter to one-third as high in humans. Taken together, the results suggest that evolutionary changes in enhancers controlling GAG-related genes lowered GAG levels in human cartilage. This molecular shift may have contributed to the distinctive features of our skeleton, from its proportions to its mechanical properties, while also helping to explain why conditions such as osteoarthritis are far more common in humans than in our ape relatives.

The significance of the work extends well beyond the skeleton itself. By combining MPRA with interspecies hybrid cells, the study allows researchers to move beyond passive genome sequence comparisons and directly test how human-specific variants affect gene activity in living cells. The same integrative approach could now be applied to investigate the regulatory basis of other traits that distinguish humans, including features of the brain and skin, and to examine how evolutionary changes contribute to diseases that disproportionately affect our species.

First author Yizhi Yan captured the scope of the achievement in a statement accompanying the release:

The study’s design reflects a broader shift in evolutionary genetics. For decades, researchers could identify candidate human-specific DNA sequences by comparative genomics, but assigning function to those sequences required slow, gene-by-gene experiments. The combination of massively parallel reporter assays with interspecies hybrid cells changes that calculus. An MPRA compresses what would once have been thousands of individual reporter experiments into a single pooled screen, with each synthetic DNA variant carrying a unique molecular barcode that allows its regulatory strength to be quantified. By testing both the human and reconstructed ancestral versions of each sequence side by side, the Kyoto University team could ask not merely whether a sequence differs between species, but whether that difference changes how strongly a gene is driven in cartilage-forming cells.

The hybrid cell strategy tackles a separate and equally persistent problem. When gene activity is compared across species in separate cultures, differences in cell type composition, growth conditions, and batch effects can masquerade as evolutionary differences. Fusing human and ape iPS cells removes much of that ambiguity, because the two species’ chromosomes share the same nucleus and the same cellular environment. Any consistent difference in expression between the human and gorilla copies of a gene within a single hybrid cell is therefore far more likely to reflect cis-regulatory changes—alterations in the DNA sequences adjacent to the gene itself—rather than differences in the shared trans-acting environment of transcription factors and signaling molecules. This distinction between cis and trans effects has been a central theme in comparative genomics, and the study’s ability to resolve it at scale is one of its methodological strengths.

The focus on glycosaminoglycans is biologically telling. GAGs such as aggrecan, the large proteoglycan that gives articular cartilage its compressive resilience, carry negatively charged sulfate groups that draw water into the extracellular matrix. This hydration is what allows cartilage to withstand the repeated loading imposed by walking, running, and jumping. A reduction in GAG content to roughly one-quarter to one-third of ape levels implies a substantially altered mechanical environment within human joints. Lower GAG abundance could change how cartilage distributes load, how it deforms under stress, and how it repairs itself over a lifetime of bipedal locomotion. The finding that GAG levels vary little among chimpanzees, bonobos, gorillas, and orangutans, yet drop sharply in humans, suggests that the regulatory shift occurred specifically on the human lineage after it diverged from the ancestors shared with the other great apes.

This result also reframes a long-standing clinical observation. Osteoarthritis is among the most common degenerative conditions in modern human populations, and it has often been attributed to mechanical factors such as our obligate bipedal gait, increased longevity, and body mass. The new data suggest an additional, evolutionary dimension: the very regulatory changes that may have reshaped the human skeleton during its evolution could have left our cartilage matrix intrinsically different from that of our ape relatives, potentially predisposing us to joint degeneration. If confirmed, this perspective could inform how researchers think about cartilage biology and joint disease, treating some aspects of human susceptibility not as pathology alone but as a byproduct of adaptations that were, on balance, evolutionarily successful.

The scale of the underlying comparative dataset is also worth appreciating. Screening roughly 5.7 million human-lineage-specific changes and narrowing them to about 561,000 candidate regulatory variants, and then to approximately 15,000 variants with measurable functional effects, illustrates how layered filtering can convert an overwhelming catalogue of sequence differences into a tractable set of testable hypotheses. Each filtering step—comparative genomics, annotation of promoters and enhancers, and functional reporter screening—discards candidates that may be real sequence differences but lack detectable effects on gene activity in the cell type tested. This also means the atlas is necessarily incomplete: variants that matter only in other tissues, at other developmental stages, or through mechanisms not captured by reporter assays would not appear in the current map.

That caveat points to the study’s most promising future direction. The skeleton is only one of the systems in which humans diverge from other apes, and the same pipeline—comparative variant discovery, MPRA screening in relevant cell types, and validation in interspecies hybrid cells—could be applied to neural progenitors for brain-related traits, or to keratinocytes and melanocytes for skin. The authors’ own framing suggests that the atlas is best understood as a proof of concept: a demonstration that the regulatory changes underlying human uniqueness can be systematically identified, functionally tested, and connected to specific molecular pathways. As such datasets accumulate across tissues, they may gradually assemble a functional picture of how the roughly one to four percent of our genome that differs from chimpanzees is deployed, and why that deployment produced a lineage that walks upright, builds large brains, and experiences a distinctive set of evolutionary trade-offs in its joints, bones, and beyond.

Subject of Research: Functional atlas of human-specific regulatory variants that shaped skeletal evolution

Article Title: How evolution shaped our skeleton: hidden clues in the genome

Article References: How evolution shaped our skeleton: hidden clues in the genome. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: human evolution, genome, regulatory DNA, enhancers, cartilage, glycosaminoglycans, MPRA, iPS cells, interspecies hybrid cells, skeleton, osteoarthritis, Kyoto University